Deep Learning-Based NOMA System for Enhancement of 5G Networks: A Review.

Abstract

The fresh and rising demands for high-reliability and ultrahigh-capacity wireless communication have led to extensive research into 5G communications. The wide progress in deep learning (DL) and nonorthogonal multiple access (NOMA) technologies provides countless benefits to communication systems. This survey provides the broad scope of DL-based NOMA for the augmentation of 5G networks. It explores various works conducted on NOMA, DL, and their applications in 5G communication. This article further explains the prominence of DL in NOMA and reviews various DL-supported NOMA models exploited for distinct tasks including resource allotment, power allotment, subchannel/channel allotment, signal detection, user detection, and other purposes. It explains the advantages and shortcomings of applying DL to resolve NOMA challenges and also presents the cardinal use cases of NOMA. It then presents a tabulated comparison of diverse DL techniques adopted by available studies for performing heterogeneous operations in NOMA. Finally, this article investigates diverse, significant technical hindrances prevailing in DL-based NOMA systems and in the application of such systems toward 5G enhancement along with future directions for shedding light on paramount developments required in existing systems for invigorating more research and supporting the emerging applications in this field.